'Singapore likely to be world’s first market for self-driving cars'

Karl Iagnemma is the CEO of nuTonomy (the company behind the deployment of self-driving taxis in Singapore last month) and also a Principal Research Scientist at MIT. He spoke to Narayan Lakshman about the potential of autonomous vehicles to transform the future of the transportation industry worldwide. Edited transcript:

How did you decide to get into this business, especially when titans such as Google, Uber and Tesla were already in the game?

My co-founder [Emilio Frazzoli] and I have been in this space for our entire careers. We come from an academic background. Both of us were directing research labs at MIT [Massachusetts Institute of Technology] for about ten years, focused on robotics technology. All of the technology going into self-driving cars is coming from the robotics community. So it was really a natural step for us to form nuTonomy because it gave us a chance to transition a lot of the research that we had been doing at MIT to the real world.

In fact, a lot of the traditional automakers are having a bit of trouble developing the technology in-house because in part the technology is not traditional automotive technology but rather robotics technology.

So would you describe it as an overlay of robotics above automotive technology?

It is really robotics technology from the ground-up in the sense that the centralised algorithms that are controlling the motion of the car, the perception of the car, and the decision-making of the car come from the robotics and machine vision communities. It is really quite new to the automotive space and that is partly why you are seeing such rise of new entrants in this space. There is opportunity to bring to bear new methods and new technologies to this automotive space and by doing that create this new product with a massive market opportunity.

Why did you pick Singapore? Was it seen as an ideal market or testing ground from which you could fan out further?

Singapore is likely to be the world’s first market for self-driving cars. Commercial services around self-driving vehicles will likely be in Singapore before they are anywhere else in the world. There are a number of reasons for this.

The main reason is that the Singaporean government has realised that self-driving vehicles could have a significant positive impact on the economy, the transportation efficiency, and on the public health and safety of transport in Singapore. The regulatory environment is very capable, the infrastructure is very good, the weather is favourable for testing and developing the technology, there are good driving practices and adherence to driving rules in Singapore.

You put all of this together and it makes it a great market for testing and development of the technology, and then eventually for the deployment of the service.

As you said that I was thinking that in India driving conditions are the exact opposite of the adjectives you used there. Is there any scope for India to consider, in the considerable future, the applications of this technology? On roads or elsewhere?

There definitely is. The reason is because the way that the technology will enter the market is first through a tightly-bounded geographical region, or in specific use cases. For example, it may be that we first see self-driving cars on the road doing things like transport between an airport and a central business district in a city.

You mentioned India as an example. In India I know that we can find regions where there is good infrastructure and there are favourable operating conditions for these cars. Today these areas may be pretty small geographically but over time they will expand out.

Overnight you will not see thousands of self-driving cars in New Delhi because the technology today is not sophisticated enough to handle driving there, it would not work so well. But in the future, sure, although it will be a measured pace of deployment in places like India.

What do you say to folks who ask you about whether, owing to the nature of this technology, you are disrupting the labour market, substituting capital or technology for labour? Is that a concern anyone has brought to you?

In Singapore it is interesting because there is concern that the taxi and bus driver fleets are ageing populations and also there is a lack of workers to actually work during some of the off-peak hours and less desirable working hours. So in Singapore this is actually that has the potential to fill a labour gap.

More broadly, that is a good point, that any automation or artificial intelligence technology will potentially have an impact on the labour market. That is something that as a society we are going to have to figure out how to reckon with.

For self-driving cars in particular there are lots of dimensions of this technology beyond just impact on labour. There is a massive potential improvement in public health. In the U.S. alone there are 35,000 deaths due to traffic accidents per year and this technology has the potential to really reduce that number.

It is a complex space but in my opinion a massive net benefit when you look at all the factors.

Did you not consider the U.S. as a potential testing ground?

We are quite interested in testing in the U.S. In past the market in the U.S. has been a little more difficult to predict because the regulatory environment is evolving, so it is a work in progress. We have tested in the U.S. with partners. We have done extensive testing in Michigan and we plan to do more extensive testing in the fairly-near future.

Do you see yourselves as succeeding against the likes of Tesla, Google and Uber in terms of market share, as all these companies roll out their product?

Absolutely. There are a lot of markets for this technology worldwide. What we are going to see is that there will be competition regionally, there will be players within different geographies who will compete and there will likely be a small handful of winners in each geographic region.

That whole landscape is still evolving and is hard to predict. Groups like Uber and Google will be looking to gain a foothold in different parts of the world in coming years.

Looking at the AI itself, there was in the past at least one case of an accident with a self-driving car of Google. How sophisticated is this AI and how long will it be before it reaches a point where you are ready to simply roll it out?

The software is being developed by each team independently. Broadly speaking there is fast progress in technology development across the industry. I think that has been proven by the fact that a small number of companies are confident enough in their software to be able to put it on public roads, like nuTonomy.

The software will have to be very, very good, high-performance, stable and robust before you can put it on the road as a product. Even when the software is at that state there will unfortunately be accidents from time to time. Some of those will not be the fault of the vehicle or the software but some probably will be, because of failures of the hardware or software, unpredictable events and so on.

One of the conversations that will have to be taking place – and it was sparked by the Google accident – is what will happen in cases where there is an accident with an autonomous car? What are the legal and liability implications of those accidents and how do we feel as a society about accidents like that?

This is all stuff that is hard to predict. It is difficult to understand the general public’s reaction to these scenarios until they actually happen. I would say that it is a ‘known unknown’ in the community but it is something that everyone is just going to have to work through.

How, for example, is Singapore dealing with this question of legal liability for accidents?

They do not have, in Singapore, firm regulation around the liability aspects today. We are confident that we will find a welcoming environment in Singapore but the specifics are all a work in progress.

Our code of editorial values

Related Topics
This article is closed for comments.
Please Email the Editor

Printable version | Dec 7, 2021 4:22:27 PM |

Next Story